Project Summary A major therapeutic obstacle in advanced colorectal cancer (CRC) is tumor recurrence at metastatic sites after treatment. These resistant and metastatic cells are thought to emerge from ?cancer stem cells?, also termed tumor-initiating cells (TICs), which are rare populations of cancer cells with stem cell characteristics. In the normal colon, multiple stem cell populations exist in a reserve-to-active continuum. We hypothesize that analogous TIC populations exist in CRC, and different populations of TICs have different clonogenic properties, tumor-repopulating potential, and propensities to metastasize. Using Lrig1+ and Lgr5+ TICs to represent reserve and active TICs, respectively, we will use novel single-cell approaches on a novel metastatic model of CRC to investigate the behaviors of TIC populations in the context of therapeutic intervention. First, we will determine whether there is a defined directional hierarchy where reserve TICs give rise to active TICs. Second, we will determine whether TICs with malignant characteristics pre-exist in the primary tumor or if TICs acquire these characteristics de novo as a result of treatment, with observed endpoints of tumor burden and metastasis after treatment. Third, we will determine whether a signature derived from TIC behaviors provides prognostic and/or predictive information for CRC metastatic recurrence in patients. Our goal is to apply a systems level strategy to predict metastatic recurrence based on single-cell properties of TICs, with the long- term goal of designing interventions for targeting TIC behaviors.